CleverClogs

Animal with me while I get to know this bear

Archive for October, 2006

links for 2006-10-19

without comments

Written by CleverClogs

October 19th, 2006 at 9:29 pm

Posted in Uncategorized

links for 2006-10-18

without comments

Written by CleverClogs

October 18th, 2006 at 7:25 pm

Posted in RSS

OPML Reader: OPML Auto-Discovery Extension for Firefox

with 2 comments

Sergio Longoni just informed me that it’s now safe to announce, review and discuss OPML Reader, an OPML Auto-discovery Extension for Firefox: this plug-in positions itself in your Firefox status bar as a tiny OPML icon. As soon as a user opens a website that links to an OPML file, the icon turns blue. Sites that do not link to an OPML file make the icon turn grey. Clicking on the OPML icon allows you to display the OPML file in Grazr or using the Optimal OPML browser. You can also opt to download the OPML file to your local hard drive.

Opml_reader_status_bar_icon

I tested Sergio’s extension and confirm that it works fine on Firefox versions 1.5 and 2.0. Note that as usual you need to restart Firefox before the extension works. For those who understand Italian, I suggest that you read Sergio’s own blog post OPML Reader per Firefox.

About Sergio Longoni
Sergio ‘Kromeboy’ is a prolific Italian geek blogger who recently provided Sterling Camden with constructive feedback on Sterling’s OPML blogroll widget for WordPress
(original conversation on James Corbett’s post OPML Autodiscovery), so it made perfect sense for me to connect to Sergio a few days ago. We exchanged quite a few messages over Skype and I’m proud that my bookmarklet inspired him to create this nifty extension. Some Firefox users happen to prefer extensions over bookmarklets so this is a perfect complement to my earlier efforts. I
know Sergio is eager to finetune his extension. Contact him directly by
looking for his ‘Contattami’ details on his blog KromeBlog if you’ve got some constructive feedback to share, or scroll down here to the comments section.

Making your site "OPML auto-discoverable"
It’s quite easy to add a link to an OPML file to your website, as long as you have access to your website’s template or source code. Go to the <head> section and insert the following (x)html tag:

<link rel="outline" type="text/x-opml" 
title
="Title of Your OPML File"
href="http://www.whatever.com/youropml.opml" />

so, in my case, my site’s source code contains this tag:

<link rel="outline" type="text/x-opml" 
title="Marjolein's Writings"
href="http://www.blogbridge.com/rl/2417/
Marjolein%27s+Writings.opml"
/>

(Split across lines because of column width)

Update 2006-10-27: By Randy Morin’s request, after a lively discussion over on Tom Raftery’s I.T. views blog between Randy and Tom Morris, I’ve changed the type attribute to "text/x+opml".

Update 2006-10-28: The debate on the proper content of the type attribute isn’t over yet. See James’ Holderness’ contribution over at Randy’s post Understanding Auto Discovery. Relying on Randy’s good judgment I changed all occurrences of type to "text/x-opml".


Suggestions for improvement

Web pages can link to multiple OPML files. At this moment the OPML Reader extension is only capable of displaying a single OPML file: it picks the last OPML file that was linked to from the <link> tag in the page header. In the case of CleverClogs this is my Marjolein’s Writings OPML, which is fine, but I can imagine other people usually list their OPML files in order of importance. So, until the extension features an OPML file selector (hint, Sergio!), I suggest that the extension selects the first one from the list.

Opml_reader_options_dialog

Other OPML Viewers

New OPML viewers are bound to be launched anytime soon. In addition to the ones I mentioned (Grazr and Optimal), there’s also Bitty Browser, for example. I have used Bitty many times on CleverClogs in the past to display websites and RSS feeds in several of my blog posts. This week Scott Matthews, developer of Bitty, pointed out to me in a private email exchange that Bitty now also supports the display of OPML files. You can try Bitty Browser here: Marjolein’s Writings Bitty-wise
 

It would be nice if the OPML Reader extension would support Bitty Browser and any other OPML viewer that a Firefox user might prefer, for example by using the same mechanism that is used for connecting applications to file extensions: open this file using %1.

I hope Sergio’s efforts also bring us closer to a solution to a problem I brought up before: when clicking on any OPML hyperlink or icon, I would like to open that OPML file in my preferred browser. Please use the comments section to this post if you have ideas about this.

Written by CleverClogs

October 18th, 2006 at 1:08 pm

links for 2006-10-16

without comments

Written by CleverClogs

October 16th, 2006 at 3:19 am

Posted in RSS

Skype RSS Reader: the Luobotou/anothR.com RSS Robot

with 2 comments

Chinese programmers have made a Skype robot available that allows you to subscribe to RSS feeds through the familiar Skype chat interface. Their website Skype RSS Robot (updated 2006-10-19) is available in Chinese and in English. Using the bot is incredibly simple: add the anothR.com bot to your contact list (the original screenname Luobotou2.0 was deprecated 2006-10-18, mh) and send it the urls to the feeds to which you want to subscribe.

Skype_bot_luobotou_1

Auto-discovery still loudly buzzing in my head, I checked if Skype RSS Robot could also detect any of my feeds by just providing it the url to CleverClogs:

Luobotou_autodiscovery_1

I’ll post an update here to see if Skype RSS Robot indeed is able to notify me of the CleverClogs blog being updated with a blog post about ‘Skype RSS Robot’ (don’t you love those Escherian deadlocks). (Update 2006-10-19: the bot finally shows my feed items)

The Skype RSS Robot page also offers RSS publishers the option to generate a subscription chicklet, but I wasn’t able to get it to work with my FeedBurner feed. (Update 2006-10-18: the chicklet finally works for me: )

From the website design Skype RSS Robot seems very early stage ("BOLG" instead of "BLOG", and even that link leads nowhere), but the main functionality definitely is in place. (Update 2006-10-18: and another improvement here: there is now a Robot’s Blog, which even mentions that you can use the Skype bot to validate your feeds, Just dump them in the Skype chat window and if the bot reports that your subscription was successful, then the validation succeeded also)

UPDATE: 2006-10-17 Open-source evangelist Chris ‘Joe Factory’ Messina of Citizen Agency found my blog post and discovered which commands are supported by Skype RSS Robot. Most important feature that I had been wondering about is that you can schedule how often the robot delivers news items. Read Chris’ post on his blog "FactoryCity": Luobotou RSS Robot — feeds for Skype.

Written by CleverClogs

October 16th, 2006 at 12:11 am

OPML Auto-discovery Bookmarklet

with 9 comments

"The idea to create an OPML autodiscovery bookmarklet came up last Friday in a very inspiring conversation with James Corbett.
I started the IM chat by asking James if he "had a minute", and if he knew of a way to make
hyperlinks to OPML files on anyone’s website to automatically open in
Grazr. After a while it turned out that for months James had been
thinking along slightly different, but converging lines: wouldn’t it be cool if OPML
files could be auto-discovered, similar to what Firefox has been doing
for RSS feeds?
"

I’d like to present my first, serious attempt to create a browser add-on that hopefully will prove to be useful to some: an OPML Auto-discovery Bookmarklet that displays a list of hyperlinks to the OPML files that have been made available by the author of any web page. 

So, what is auto-discovery anyway? Most Firefox users are familiar with the orange feed icon to the right of the address bar, indicating that there’s an RSS feed attached to the page they are viewing. This mechanism is called auto-discovery. There’s not much magic about it: web page authors deliberately enable auto-discovery by linking to feed urls with an (x)html tag named <link> in the source code of their pages. Auto-discovery of OPML files is based on the same principle: detecting <link> tags in the header of a web page.

Opml_autodiscovery_bookmarklet

The OPML Auto-discovery Bookmarklet scans the header of a website and displays a list of OPML outline files in a box in the top-left corner of the browser. For completeness’ sake any RSS feeds are listed too. Because of the debate about the proper contents of the "rel" and "type" attributes for outline files I decided to be liberal what to accept: any "href"attribute that contains the text "opml" is accepted. (Update 2006-10-18: I slightly altered the code to also catch those OPML files that use a file extension of ‘.xml’  and have their type attribute set to ‘text/xml+opml’ or ‘text/x-opml’)

Below is a screenshot of what the CleverClogs header looks like, for example:

Cleverclogs_page_header

The bookmarklet lists each of these files as Grazr hyperlinks: when you click on any of these a Grazr window is opened in a new tab, displaying the contents of the OPML file or the RSS feed. Note that if a "title" attribute is provided, the bookmarklet will take that. Otherwise the url to the file is used to populate the box.

To install the bookmarklet drag this link OPML Auto-discovery Bookmarklet (updated 2006-10-18) to your Bookmarks Toolbar. You can name the bookmarklet as you wish, but I suggest to use "OPML Discoverer". The source code is also available for download as txt files: see the references at the bottom of this post.

The idea for this project came up last Friday in a very inspiring conversation with James Corbett.
I started the IM chat by asking James if he "had a minute", and if he knew of a way to make
hyperlinks to OPML files on anyone’s website to automatically open in
Grazr. After a while it turned out that for months James had been
thinking along slightly different, but converging lines: wouldn’t it be cool if OPML
files could be auto-discovered, similar to what Firefox has been doing
for RSS feeds?

What followed was a lively chat session spread out through the day. James passed on links and code snippets to me, we determined the essential goal of the bookmarklet and then I started to work on the script. 

Lately quite a few other people have posted too about their desire to have OPML files auto-discovered, as you can see in this highly relevant, live blog search panel based on a simple Google blog search:

Grazr panel

My increased interest in OPML is a natural consequence from my
newsmastering experiments. The biggest boost occurred when I started
developing RSSonate, a list of RSS feeds about RSS technology implemented in Grazr. A few weeks after I started it RSSonate was turned into a BlogBridge expert library. RSSonate is still on live display in the sidebar of CleverClogs, under the heading ‘Others Wrote’.

Note that I didn’t create this bookmarklet from scratch: I adapted
existing source code for a similar bookmarklet made available by
Google—credit to whom credit is due. The Google "Show All Feeds"
bookmarklet was originally announced by Chris Wetherell in the blog
post Subscribing to feeds via little Google buttons
on the Official Google Reader Blog in November 2005. Since then several
others have contributed to the source code, most notably Martin Dittus with Bookmarklet: Display Feed Links on Current Page (March 2006).

My adaptations consist of
showing hyperlinks to OPML files and some cosmetic changes, like
displaying the blue and orange icons and a link back to this CleverClogs post.
Although I do have some programming background, bluntly speaking Javascript programming isn’t my specialty at all. Both James Corbett and I believe there’s room for improvement, so let’s foremost consider this a proof of concept. Hence I kindly invite Javascript programmers
to look at the underlying code and make constructive suggestions for improvements. If you blog about this project, then please send a trackback ping.

Making your site "OPML auto-discoverable"
It’s quite easy to add a link to an OPML file to your website, as long as you have access to your website’s template or source code. Go to the <head> section and insert the following (x)html tag:

<link rel="outline" type="text/x-opml" 
title
="Title of Your OPML File"
href="http://www.whatever.com/youropml.opml" />

so, in my case, my site’s source code contains this tag:

<link rel="outline" type="text/x-opml" 
title="Marjolein's Writings"
href="http://www.blogbridge.com/rl/2417/
Marjolein%27s+Writings.opml"
/>

(Split across lines because of column width)

Update 2006-10-27: By Randy Morin’s request, after a lively discussion over on Tom Raftery’s I.T. views blog between Randy and Tom Morris, I’ve changed the type attribute to "text/x+opml".

Update 2006-10-28: The debate on the proper content of the type attribute isn’t over yet. See James’ Holderness’ contribution over at Randy’s post Understanding Auto-Discovery. Relying on Randy’s good judgment I changed all occurrences of type to "text/x-opml".

Sites that have added an OPML file to their page headers:


Source code links:
OPML Auto-discovery Bookmarklet, formatted source code (updated 2006-10-18 – 11:50 PM)
OPML Auto-discovery Bookmarklet, unformatted source code(updated 2006-10-18 – 11:50 PM)

P.S. The ironic part of this weekend’s marathon project is that I solved a problem that wasn’t my own. Now, is there anyone who knows how to solve my original problem: is it possible to instruct Firefox to display any OPML file using Grazr, in the same way that PDF files are displayed using Adobe Acrobat?

Updates:

2006-10-16, 02:56 AM – TailRank lists all posts linking to this story.

2006-10-16, 10:46 AM – David Rothman submitted a digg about my bookmarklet. If you like, go and cast a vote.

2006-10-16, 10:50 AM – I slightly altered the script so that it only
detects the string "OPML" when it’s at the end of the href attribute.

2006-10-16, 11:28 AM – Sergio Longoni informed me that he is working on a Firefox extension that will display a blue OPML auto-discovery icon in the address bar.

2006-10-16, 11:50 AM – Tom Morris blogs "I’m going to write a PHP API to do similarly on the server side – you
provide a URL and it returns an OPML 2.0 file containing links to all
the OPML and RSS feeds
"

2006-10-16, 12:44 PM – It would be great if the search results pages generated by Andy Edmonds’  just released Live Search OPML Generator would also have a <link> tag in their page headers, so that this bookmarklet and other Firefox add-ons can detect them (I tried to post a comment to Andy’s MSDN blog post and to his Surf*Mind*Musings blog post, but they both require that commenters be signed in—too bad, trackbacking instead.)

Written by CleverClogs

October 15th, 2006 at 4:23 pm

links for 2006-10-14

with one comment

Written by CleverClogs

October 14th, 2006 at 3:25 am

Posted in RSS

links for 2006-10-10

without comments

Written by CleverClogs

October 10th, 2006 at 11:21 am

Posted in RSS

links for 2006-10-10

without comments

Written by CleverClogs

October 10th, 2006 at 7:31 am

Posted in RSS

links for 2006-10-10

without comments

Written by CleverClogs

October 10th, 2006 at 3:23 am

Posted in RSS